One $8 million liability verdict lands in a book that averages $400,000 of loss per year. Leave it in the data and next year's rate doubles for everyone. The fix is not to ignore the loss but to spread it.
A rate indication assumes the historical loss experience predicts the future policy period. That works when losses are frequent and stable. It fails when a single event dwarfs a normal year. Three sources cause this: individual large losses (a huge verdict or shock claim), catastrophes (hurricanes, wildfires, hail, correlated events hitting many policies at once), and pandemic effects (a temporary systemic shift in frequency or severity).
Each shares one property. The event is rare, so it appears in some experience years and not others. Averaging a few years of raw data either overstates the rate (if a shock landed in-period) or understates it (if the quiet years happened to be clean). The goal is to base the indication on the expected cost of these events, not the realized cost of a short window.
Common mistakes
- Deleting the large loss outright. Capping keeps the first $1,000,000 and reloads the excess as an average. Removing the entire $2,500,000 claim understates expected cost.
- Capping untrended losses at a fixed threshold. Inflation drifts claims over a fixed cap, so an excess ratio measured on untrended losses is too small and the excess loading comes in light.
- Applying a total-limits trend to basic-limits losses. The basic layer trends slower than the total (4.35% versus 10.0% in Example 3). Adjust a total-limits trend down before using it on capped losses.
Bottom line
- Large losses, catastrophes, and pandemic effects are real but volatile; left raw, one year's shock distorts the whole indication.
- Cap individual losses at a large-loss threshold chosen by balancing loss volume against volatility, pool the excess across years, and reload it as a smooth loading.
- Trend losses to future levels before capping (or index the threshold); a fixed cap on untrended losses understates the excess ratio and the excess loading.
- Limits leverage trend: basic-limits trend at most total-limits trend at most excess trend for positive trend, reversed for negative trend.
Exam shortcut
When a claim exceeds a stated cap, split it immediately: keep the amount up to the threshold in the data, and reload the excess using the given long-run ratio times the capped losses. Never touch the raw total. When trend is in play, trend the ground-up losses first, then cap.
The full lesson (about 3,159 words, 21 min read) adds 4 worked examples, all 8 common mistakes, a self-check, free in the app.
Learning objectives
- A10
Browse all free Exam 5 lessons or jump into free Exam 5 practice questions.